Must-read for agent operations: Guide to deduplication, acceptance and archiving of tg US data by project in multi-client scenarios

When the agency operation team manages Telegram community customer acquisition projects for multiple overseas customers at the same time, the reuse and isolation of tg US data is a core pain point. Mixing the same number pool will not only cause customer A’s number to be repeatedly charged by customer B’s number screening task, but may also cause privacy disputes due to data pollution. This guide is intended for agent operations and cross-border customer acquisition studios. It systematically explains how to use the KK-DATA platform to efficiently complete the deduplication, acceptance and archiving of tg US Data sub-projects in multi-client scenarios, thereby increasing per capita production capacity and reducing balance waste.

Why does the agent operation team need to pay attention to the sub-project isolation of “tg US data”?

Agent operations usually serve 3–5 or more customers at the same time, and each customer may target a different American population (for example: young women in New York State, technology practitioners in California). If project isolation is not done, three serious problems will occur:

• Data Pollution: Customer A’s invalid number was misjudged as the target number, and Customer B’s screening task was imported, resulting in deviations in subsequent marketing results.
• Duplicate Charges: The same number is detected multiple times in different customer tasks, resulting in unnecessary consumption. Taking 10,000 numbers as an example, if the repetition rate reaches 20%, approximately 2,000 detection fees will be wasted (see the real-time price on the console for details).
• Acceptance Disputes: The source of a certain batch of data cannot be traced, and it is difficult for customers to self-certify when they question the quality of the data.

The core value of sub-project isolation is to make every piece of tg US data traceable, auditable, and reusable. This is not the icing on the cake, but the bottom line for large-scale delivery by the operations team.

How to use deduplication warehouse to achieve zero duplication of TG US data across projects?

KK-DATA’s data deduplication warehouse is an automatic deduplication mechanism at the account level. When you submit a new task, the system will compare all the numbers that have been detected by the current account and automatically skip duplicates without being billed or appearing in the results. This means that the same number will only be detected once within this account, no matter how many tasks it appears in.

Setting up steps and best practices for deduplication warehouse

1. Log in to the console → Enter the “Deduplication Warehouse” page (enabled by default, no additional configuration required).
2. Before submitting the task: Make sure that the list of numbers to be detected contains numbers that may be duplicates. Just upload them directly and the system will automatically filter them.
3. Best Practices:
   • Create a new independent task for each customer, and mark the customer abbreviation and batch in the task notes (such as ClientA_US_tg_batch1).
   • Do not mix multiple customer numbers into the same document submission to avoid follow-up difficulties.
   • Export deduplication warehouse status reports regularly (such as every week) to check inspection records with customers.

Quantitative estimate of balance savings due to deduplication (taking 10,000 tg US data as an example)

Suppose your agent operation team holds 10,000 pieces of tg US data, which are repeatedly referenced by 3 different customer tasks within three months:

• No deduplication warehouse: 3 tasks detect 10,000 items each → 30,000 items fee (deducted based on actual unit price).
• Use the deduplication warehouse: the first task detects all 10,000 items, and the latter two tasks only detect the newly added parts (assuming that the latter two tasks include some known numbers and do not introduce new numbers themselves), the actual fee is about 10,000 items.

After the initial inspection, subsequent tasks cost almost nothing. Even if each customer has independent number requirements, as long as the full amount of data is not deliberately mixed into the same task, the deduplication warehouse can effectively intercept duplication. For more accurate cost savings, please refer to the task estimated cost interface of the console.

In a multi-client scenario, how to set the naming convention and project coding for tg US data?

Standard file naming allows you to locate the specified customer’s data within 30 seconds after exporting the CSV/TXT results. The following format is recommended:

[客户缩写]_[地区]_[数据批次]_[日期]_[状态].csv

Example:

• CA_CA_tg_vol1_20250401_active.csv (Customer A California tg active data)
• CB_NY_tg_vol2_20250405_gender.csv(Customer B New York tg gender data)

Project management method in KK-DATA console

When submitting a screening task, you can use the Task Notes field (located below the number upload area) to enter the customer initials and batch. The file name of the exported result file can be customized, and it is recommended to follow the above format. In addition, the console provides a “Historical Tasks” list, which can be filtered by time, notes, and status. It is recommended to add notes immediately after each task is completed to avoid forgetting.

Commonly used fields list (recommended to keep when exporting custom fields):

Field	Description	Required
phone	number	yes
tg_status	Activated/Not activated	Yes
tg_active	Activity level (such as active within 7 days)	According to customer needs
gender	gender	according to customer needs
age	Age (range)	According to customer needs
task_id	Task ID	Recommended to be reserved for acceptance
remark	Customer remarks	Fill in manually

Acceptance process: How to confirm that the quality of tg US data meets customer requirements?

Acceptance is the core of agency operation service delivery. It is recommended to formulate acceptance standards from four dimensions:

1. Opening rate: refers to the proportion of detected TG registered numbers. Industry normal value: The TG activation rate in the United States is usually between 60%–80% (affected by the source of the number segment). If it’s less than 40%, you need to recheck the number source.
2. Activity: Customers usually require active within 7 days or 30 days. The activity rate generally accounts for 50%–70% of the opened numbers.
3. Gender/Age Distribution: If the customer targets males aged 25–35, the number of numbers in this age group can be required to account for no less than 30% of the total active numbers.
4. Field Completeness: The missing rate of fields such as gender, age, avatar, etc. should be less than 5%.

What key indicators should the acceptance report include?

It is recommended to generate the following simple table (example):

Indicators	Actual values	Customer requirements	Whether standards are met
Total Number	10,000	≥10,000	Yes
Number of TG activations	7,200	—	—
Opening rate	72%	≥65%	Yes
Active count (7 days)	4,680	—	—
Activity rate	65%	≥60%	Yes
Male proportion	58%	≥50%	Yes
Proportion of people aged 25–35	35%	≥30%	Yes
Field missing rate	2%	≤5%	Yes

Each indicator can be accompanied by a text explanation, for example: “The activity rate is based on login behavior within 7 days. If the customer needs a closer window, they can re-test.”

Common reasons for acceptance failure and quick fixes

• The activation rate is too low: Check whether the number source is a cleaned number segment; use the global number generation module to randomly generate a US number segment (can be generated for free), and then screen again.
• Activity rate is low: It may be that the detection window is too long or short and the default settings are unreasonable. When resubmitting the task, manually select “Active within 7 days” or “Active within 30 days”.
• Gender ratio is not met: If the proportion of males is too high, you can add a gender filter when filtering numbers to filter female numbers separately.
• Multiple missing fields: Some numbers may be virtual operators or zombie numbers, and the missing information is normal. The sample size can be increased to filter out numbers smaller than a certain threshold.

Archiving Policy: What information should be retained after the tg US data project ends?

After the completion of each project, it is recommended to retain the following documents for a minimum period of 30 days (can be extended to 90 days based on contractual requirements):

• Original number file: CSV file before filtering (including source notes).
• Screen Result Report: Full CSV/TXT exported from the console, including all sieve fields.
• Deduplication Record: Task logs can be exported to mark which numbers were skipped from the deduplication warehouse.
• Acceptance Report: Contains sampling data, customer confirmation signature or screenshot.
• Customer Feedback: Records of any communication regarding data quality.

Archiving operation example: Store the above files locally in the customer folder (such as ClientA_TG_US_2025Q1), and the file names must contain the date. At the same time, save task notes in “Historical Tasks” in the KK-DATA console. It is recommended to add “Archival Date” to the notes. In this way, even if the local file is lost, the results can be re-downloaded through the console (note: the download time limit is subject to the platform storage policy, local backup is recommended).

Workflow summary of the operation team using KK-DATA to manage tg US data

The following is a typical SOP for completing a multi-client tgUS data task within one day:

Steps	Time estimate	Operation content
1. Requirements sorting	15 minutes	Collect each customer’s TG filter requirements (region, active window, gender, etc.) and organize them into a task list
2. Number generation	30 minutes	Use the global number generation module to generate US number segments according to customer requirements; or import CSV from an existing database
3. Deduplication check	5 minutes	Make sure the deduplication warehouse is turned on before submitting the task; if there is cross-account data, manually deduplicate it first
4. Submit the screen number	10 minutes	Create tasks for each customer in the console, fill in the notes; estimate the cost and confirm the balance
5. Wait for completion	1–6 hours	Depending on the number of numbers, Telegram notification can be set
6. Export results	10 minutes	Export CSV by customer and name as required
7. Acceptance review	30 minutes	Generate an acceptance report based on the above four dimensions and sample for confirmation
8. Delivery and archiving	15 minutes	Package files (raw data + reports) and send to customer while archiving locally

It takes about 2–8 hours in total (depending on the number of numbers and platform processing speed), and multiple customer tasks can be processed in parallel (KK-DATA supports submitting multiple tasks at the same time without interfering with each other).

FAQ

Q: Are the gender and age fields in tg US data accurate? Can it be used for precise orientation?

Answer: The gender and age fields are based on algorithm inference and are not real-name authentication data. The accuracy rate is about 80%–90% (depending on the platform). It is suitable for crowd trend analysis (such as “men around 30 years old”), but is not suitable for scenarios that require ID card-level accuracy. It is recommended to combine it with business verification before using it on a large scale.

Question: When operating multiple customers, will KK-DATA’s deduplication warehouse automatically take effect on all tasks?

Answer: Yes. The deduplication warehouse is an account-level function. When submitting a new task, the system will automatically compare the detected numbers and skip duplicates, without manual configuration. However, data between different accounts will not be interoperable. If you need to deduplicate data across accounts, you can export it and process it manually.

Question: Can a task contain the tg US numbers of multiple customers?

Answer: Technically yes, but it is strongly recommended that each customer create a separate task and indicate the customer ID in the task notes. This facilitates subsequent acceptance, archiving and balance reconciliation, and also avoids disputes caused by data confusion.

Question: After exporting tg US data, can it be imported into the deduplication warehouse again?

Answer: Yes. You can import the number column in the CSV of historical results into the “Customized Number Segment” function in the “Generate Number” module of the console, and let the system automatically remove duplicates through the submission process. However, it is more recommended to complete all operations directly in the console to reduce manual intervention.

Question: If the customer requires the original testing source certificate of tg US data, can KK-DATA provide it?

Answer: The platform provides the platform detection timestamp and task ID in the exported fields, which can be used as recording credentials. However, it is impossible to provide the underlying interface logs or original screenshots of the third-party detection platform. It is recommended that the data source be disclaimed in the service contract.

---

If you are worried about how to manage multiple customers’ tgUS data, you might as well try the above workflow today. KK-DATA provides one-stop screening, deduplication and export capabilities, helping the agency operation team save more than 30% of repeated testing costs. Try it now:

👉Log in to the console to start screening numbers Two-way contact customer service: https://t.me/kkdata_robot Official website: https://kkdata.cc/ Documentation: https://docs.kkdata.cc/